High Resolution least-squares wave equation AVA imaging: Feasibility study with a data set from the Western Canadian Sedimentary Basin

Summary This paper presents a regularized least-squares pre-stack 3-DwaveequationAmplitudeversusAngle(AVA)migrationalgorithmandexploresthefeasibilityofthisclassof methodstoprocessflelddata. Weposeseismicimagingas alinearinverseproblemthatincorporatesweightingmatricesinmodelanddataspace. Thegoalistoremoveadditivenoiseandartifactsthatarisefromdataacquisition, operatormismatch andadditive coherentandincoherent noise. We solve the inverse problem with the conjugate gradientsmethodand,inaddition,weacceleratetheconvergenceoftheCGschemebyapreconditioningstrategy. We have applied the regularized least-squares migration (RLSM) algorithm to a 3-D data set from the Western CanadianSedimentaryBasin. Theinversionsigniflcantly improvesthe quality ofthe commonimage gathers. Theaccuracyofouralgorithmisconflrmedbyadetailed comparisonofinverted andsynthetic CIGs. We also observe an substantial enhancement of vertical resolution as a consequence of improving the coherence of the inverted common image gathers and an implicit deconvolution thatis embedded in the method.